AEO for dentists: win the implant and Invisalign asks.

A hygiene cleaning and a full-arch implant case come through the same reception desk, but they are not worth the same chair — and AI answers, unlike a Google ranking, treat them as different questions with different winners. Engines compose a fresh shortlist for every ask, so a practice can own the checkup answer and be absent from the implant answer in the same city, on the same day. AEO for dentists means winning the procedure asks that fill the production ledger — implants, Invisalign, veneers — while holding the routine asks that fill the hygiene schedule, and knowing, scan by scan, whose names the engines give. This page maps the ladder, rung by rung, and what to build for each.

The procedure ladder: not all questions are worth the same chair

A Google ranking never asked what the searcher was worth to you — page one was page one, whether the click was a cleaning or a full-arch case. AI answers break that flatness. Engines compose each answer per question, which means a practice can be the name for checkups and absent for implants in the same city, on the same day. So the useful map of dental AEO is not a keyword list. It's a ladder of question types, climbing in value.

Routine asks: the standing schedule

The new-resident version is the quiet prize: a family choosing its first dentist in a new city is choosing, in most cases, its dentist for years — and new residents are exactly the people with no local friends to ask, which is why the question goes to a machine. The emergency version is the opposite tempo: no research window at all, just the first credible name with an open phone line and an answer about hours.

Invisalign and orthodontics: the researched family decision

Orthodontic asks are usually a parent's project: weeks of comparing, with cost at the center. Notice what the phrasing does — it bundles the treatment, the city, and the insurance plan into one sentence and expects one answer. The practice whose pages answer the cost and insurance parts in plain language is the practice the engine can actually use.

Implants and cosmetic work: the highest-stakes ask

A multi-thousand-dollar, irreversible procedure gets researched harder than anything else a general practice offers — these patients compare surgeons, materials, and financing for weeks before booking a single consultation. They also skew older, which used to mean offline. It no longer does: rater8, a review-management vendor serving healthcare practices, reported in its 2026 Patient Choice Report that 64% of patients aged 45 to 60 — the implant demographic — used AI to find providers.

The behavior underneath the whole ladder is measured, not guessed:

Fifteen questions, arranged by what they are worth

Below is the ladder as a chart. Every line is a question shape published in dental answer research by marketing firms working in the space — read them the way you would read tomorrow's schedule, noting which rung each one lands on and how much of the answer depends on facts your website may not state anywhere.

Three patterns run through the whole chart. Cost appears in almost every high-value question, and most practice sites answer it nowhere. Insurance appears as a filter the engine can only apply from text it can read. And the urgent rung turns entirely on hours — a fact that usually lives in an image, a widget, or a booking embed rather than in a sentence.

The one-name dynamic: there is no page two

A Google results page has ten organic spots and a three-pack on the map; position six still exists, and a determined patient scrolls. An AI answer has no positions. It typically weaves two or three practice names into a paragraph, each with a reason attached — and every other practice in the city simply does not exist in that conversation. The named practice doesn't rank first; it's the only door the patient is shown. For implant and Invisalign cases, that named-or-nothing coin flip is happening with the most valuable patients on the schedule.

How a practice loses a procedure answer

The reflex response is “we rank well on Google, so we're fine.” The two contests are separate:

The loss is not random, though. In dentistry it happens through four recognizable mechanisms:

• A directory answers instead of a practice.Ask an engine for an Invisalign dentist who takes a named plan and the most quotable page on the open web is frequently a Zocdoc, Healthgrades, or insurer find-a-dentist listing — a filtered roster that satisfies the whole question at once. The practice's own site loses to a page that merely lists it, or does not.
• Multi-location groups arrive with a page for every question. A practice group publishes a templated implant page, Invisalign page, and emergency page for each of its locations. An independent publishes one services page listing twenty-two procedures. When the engine needs a page that is about implants in your city, only one of those two exists.
• The best clinical evidence is unquotable.Reviews that say “everyone was so nice” carry no procedure and no clinician, so they cannot support an implant recommendation. “Dr. Reyes placed my implant and the follow-up was thorough” can. Most practices never ask patients for the second kind.
• Key facts are locked in formats machines skip. Accepted plans in a PDF. Hours inside a booking widget. Fee ranges in an image on the financing page. Each one is a question the engine wanted answered and had to answer from somebody else's page.

The consolation hiding in the same mechanics: nothing here is settled. 40–60% of the sources AI engines cite change every month (eMarketer, 2026), so a procedure answer that names two competitors today is contestable next month — and an answer you win needs defending on the same schedule. If your practice is on the wrong side of an answer right now, the six reasons a business stays invisible in ChatGPT is the diagnostic, and the five-prompt manual check takes about five minutes before you spend anything.

Where the engines read about dentists

Engines don't intuit that you place implants well. They assemble dental answers from evidence they can read and quote, and for dentistry that evidence sits on a recognizable stack of surfaces:

• Health directories: Healthgrades, Zocdoc, and WebMD — the profiles engines treat as third-party confirmation of who you are, what you treat, and what patients report. The ADA find-a-dentist directory anchors the credential side.
• Reviews with procedure detail: Google Business Profile and the other review surfaces, read for volume, recency, and specifics — the difference between quotable evidence tied to a procedure and a bare five-star average, which is arithmetic.
• Insurance listings:the “takes my insurance” half of so many asks gets answered from insurer find-a-dentist directories and from whether your own site states accepted plans in plain text.
• One exact practice name everywhere: website, Google profile, directories, insurance listings. Three name variants read as three weak entities instead of one strong one.
• Community threads:the city subreddits where “anyone know a good implant dentist here?” threads get quoted into answers — why AI cites Reddit so much explains the over-weighting.
• Your own pages, if the crawlers can reach them. OAI-SearchBot is the crawler tied to appearing in ChatGPT Search. GPTBot is primarily training collection, and blocking it does not remove a practice from ChatGPT Search. Anthropic runs the same split with Claude-SearchBot and ClaudeBot.

The machinery underneath every “near me” version of these questions — Google Business Profile, reviews, NAP consistency, structured data — is its own subject, covered in how AI picks local businesses. On markup specifically, keep expectations calibrated: Google states no special structured data is required to appear in AI Overviews or AI Mode. What a Dentist or MedicalBusiness block does is state hours, services, and accepted plans in a form that leaves a machine nothing to interpret — and keep the classic rich results working. The same dynamic runs one door down in aesthetics, where med spas fight national chains for the same kind of answer.

Clinical accuracy: what we will not write about your dentistry

Dental advertising sits under state dental board rules and the same truth-in-advertising standards as any medical service, and the practical risk in AI-era content is specific: a machine-assisted draft that gets a clinical detail subtly wrong, or that claims a credential the license does not carry. The useful test for any vendor is not what it promises to produce. It is what it refuses to produce. Ours:

• We will not make a clinical outcome claim.No “painless,” no “lasts a lifetime,” no success rates, no healing timelines presented as what a reader can expect. Procedure pages describe what the procedure involves and who performs it.
• We will not apply a specialty title the license does not support. Orthodontist, periodontist, oral surgeon and the other recognized specialties are regulated words in most states. A general dentist who places implants gets described as a general dentist who places implants.
• We will not invent a patient story, a review, or a result. Not as a placeholder, not as an illustration, not inside structured data.
• We will not state a plan you accept or a fee you charge unless the practice has confirmed it. Insurance text is the single most-quoted thing on a dental page and the easiest to get wrong; it comes from you, not from an inference.
• We will not put a procedure in schema that the practice does not perform — or that the visible page does not mention. Markup and page say the same thing, always.
• We will not touch patient data, ever.The whole method reads the public web: your site, your profiles, and the engines' answers. No records, no appointment data, no protected health information enters this system at any point.

And nothing publishes until someone at the practice approves it, so whichever clinician signs off on your marketing today keeps that seat. None of the above is legal or clinical advice; state board rules differ, and your own reviewer decides.

What gets built for each procedure question

Disclosure, so you can weigh this section: this page is published by AEO Action, and here is where we describe our own product. Everything above is true whoever does the work.

AEO Action runs the loop as software — daily scans that ask up to six engines your patients' questions (ChatGPT, Gemini, Perplexity, Google AI Overviews, Microsoft Copilot, and Claude, the last in a labeled model mode that shows how the model understands your practice long-term and is never counted as a live answer), with the questions regenerated each scan from your practice's profile, procedures, and city. What makes the product make sense for a dental practice is that each rung of the ladder gets its own kind of artifact:

• The urgent rung gets an hours-and-access page. Same-day availability, after-hours protocol, what to do with a lost crown before you arrive — in sentences, not in a widget. That is the only form the “open right now” question can be answered from.
• The routine rung gets a new-patient page.Written for the family that just moved: first visit, what to bring, children's appointments, which plans the practice takes, stated in plain text rather than in a downloadable form.
• Orthodontic questions get a cost-and-options page. The Scottsdale example above ends with exactly this artifact — FAQ page — “how much does Invisalign cost in Scottsdale”— a page that answers the cost question in plain language, built from the practice's own answers, with matching structured data included.
• Implant and restorative questions get their own landing page each. One question, one city: what the procedure involves, who performs it and where, what drives the cost range, whether grafting and surgery happen in-house, how financing works.
• Cosmetic and anxiety questions get a plan page. Veneers, whitening, sedation options and who administers them — the elective rung where the patient is choosing a person as much as a procedure.
• The practice itself gets structured data and a Reddit research pack. One machine-readable statement of practice, procedures, hours, and location that matches the visible pages; plus the city threads where dental questions actually get asked, prepared as research rather than as posts.

The cadence: the orchestrator picks one highest-value move per day from the scan, prepares the complete artifact, and queues it. Nothing publishes without your Approve — on day one of a trial, up to four artifacts are waiting, depending on what the scan finds. Publishing tells the truth about your website: Webflow and Squarespace connect for direct publishing; on WordPress, the plugin pulls approved schema hourly and approved drafts every six hours — we never push into your site; everywhere else you get finished HTML and schema ready to paste, and the system verifies the page went live. Pages that go live are then submitted through IndexNow within minutes, and Bing decides what to index. Why spend the step at all: independent testing found roughly 87% of ChatGPT's citations also rank in Bing's top results (Seer Interactive, 2026) — an overlap measured from the outside, worth working because it is large. OpenAI documents OAI-SearchBot as the crawler for ChatGPT Search and does not publish which index or indexes sit behind it, so nobody honest sells you a pipeline here.

Then comes the part dentists are right to demand from any marketing spend: proof. Every approved landing page is registered against its target question and re-checked on each following scan for up to 120 days — a dated receipt reading shipped on date A, first cited on date B. Published testing observes a median of about 6.8 days to a first ChatGPT citation once a page is indexed, with roughly 42% of pages cited within 30 days (Semrush / practitioner testing, 2026) — observed medians from testing, not a promise about your practice. Every seven days a plain-English report lands: what was asked, what changed, what shipped, what got cited.

The retainer question, priced honestly

Dental practices already buy marketing, so the fair comparison is the one you're living with. Full-service marketing retainers commonly sit at $2,000–6,000/month(2026 market research) — and an agency is still the right pick more often than software companies admit. If the practice needs paid search managed, mailers designed for the new subdivision, the front desk trained on phone conversion, and photography that doesn't look like stock, a retainer buys people, and people are what that job takes. The narrower case for software is a practice that wants the AI-answer loop — scan, build, publish, verify, re-check — worked every single day without adding headcount: the Solo plan is $149 a month, and what the department actually does is laid out service by service. The full price ladder, rung by rung, is in how much AEO costs in 2026, and the skeptic's version of the whole argument is is AEO worth it.

The first step costs nothing either way. The free audit asks ChatGPT, Gemini, and Perplexity ten questions patients ask about practices like yours — 30 live answers scored in about three minutes, the three most revealing shown word for word with every named competitor highlighted. The PDF is yours permanently; the share link lives seven days.